Reed 1

The Long-Term Consequences of Deer Browse in

Temperate and Boreal Forests

Honors Thesis

Samuel Powers Reed

Honors Environmental Science B.S.

Evolution, Ecology, and Organismal Biology Minor

The Ohio State University

2017

Thesis Committee:

Dr. Peter S. Curtis (Research Adviser)

Dr. Roger A. Williams (Academic Adviser)

Dr. Lauren M. Pintor

Reed 2

Abstract:

White-tailed deer (Odocoileus virginianus) overabundance has sparked dramatic changes

in forests throughout North America. In Pennsylvania’s Allegheny National Forest and Quebec’s

Anticosti Island, I investigated how deer have altered canopy structural complexity, species

diversity, and carbon stocks across a controlled density gradient and a chronosequence of

increasing disturbance, respectively. To measure canopy complexity, I used a Portable Canopy

LiDAR (PCL) system, which records the three-dimensional arrangement of leaves and stems

within a canopy using an upward-facing infrared laser. In Pennsylvania we predicted that

treatment effects on forest composition and structure would support the intermediate disturbance

hypothesis and that stands experiencing moderate levels of deer browse during the early stages of

regeneration would show increases in canopy complexity, carbon sequestration, and diversity. In

Anticosti we predicted a significant increase in canopy complexity and carbon storage between

deer preferred and non-preferred stands and as stand age increased.

Reed 3

Introduction:

Over the last century, white-tailed deer populations (Odocoileus virginianus) have

rapidly increased throughout the eastern portion of America. The increase in population size and

browse intensity has altered many forest’s regenerative capacity and their resulting structure

(Cote et al. 2004). Changes in structure and biodiversity alters a forest’s ecological functions,

such as carbon sequestration, water filtration, or wildlife habitat (Bengtesson et al. 2000). As

deer populations continue to be stressors on forest ecosystems, managers and the public must

adapt to a changing landscape or find a solution to the high population levels.

As deer change forest communities, they likely alter canopy structural complexity (CSC),

the three-dimensional arrangement of leaves and branches within a canopy. CSC has been

implicated as an important factor in aging forests’ ability to maintain functionality, such as

carbon storage or wildlife habitat (Hardiman et al. 2013b; Ishii et al. 2004, Jung et al. 2012). As

forests age and experience moderate abiotic disturbances, their canopies may become more

complex. Net primary productivity (NPP), the rate that plants take in carbon and grow, appears

to increase as a function of structural complexity, as light and nitrogen are used more efficiently

(Hardiman et al. 2013a). However, the impact of large or chronic biotic disturbances on canopy

structure have not been analyzed.

In both Pennsylvania’s Allegheny Plateau and Quebec’s Anticosti Island, I analyzed how

deer shifted forest types through selective browsing. Both sites have had high deer populations

throughout the 20th and 21st century (Leopold et al. 1947; Potvin et al. 2003) and are extensively

clearcut for timber. Clearcuts allow deer to browse on regenerating seedlings and alter canopy

structure, especially when other food is scarce during the winter months.

Allegheny Plateau:

Reed 4

There has been an overabundance of deer on the Allegheny Plateau since the early

1920’s, as hunting regulations and deer protections were established (Leopold et al. 1947).

Average deer density reached 11.6-19.2 deer/km2 in the 1930s, dropped to ~5.4 deer/km2 in the

1940s due to severe winters, and rose to 15.6-22.6 deer/km2 by the 1970s before the initiation of

this study (Horsley, 2003). Between 1950 and 1970 a renewal in timber production increased

deer densities and allowed for overabundance (Horsley, 2003).

Within the region, most herbivory studies focused on excluding deer from plots, with

some exclosures in use for over 60 years (Goetsch et al. 2011). Although exclosures provide

valuable information, there is little control over the ambient environment’s deer population or

browse intensity outside of the fencing, making it difficult to justify that the vegetative

regeneration is generalizeable. In contrast, deer enclosures allow for control over population size

and the resulting browse intensity.

In 1979 and 1980, the U.S. Forest Service established four 65 ha deer enclosures

throughout the Allegheny Plateau that were in use for 10 years (Tilghman, 1989). The four 65 ha

deer enclosures contained subenclosures of 4, 8, 15, and 25 deer/km2, which represent historical

deer densities from the late 19th century to present time. To model a 100-year heavy management

scenario, each density was treated with three different cutting schemes, clear cut, cut to 60%

residual relative density, and uncut (Tilghman et al, 1989). After 30 years, each subenclosure has

maintained the residual effects of long-term browsing, such as browse-tolerant species

dominating high deer density treatments (Nuttle et al. 2011) (Figure 2).

In this study we investigated how deer have altered canopy structural complexity,

biomass, and species diversity within each density treatment’s former clear cut (n=16). Our null

hypothesis was that deer have no significant effect on the three aforementioned variables. In

Reed 5

contrast, our alternative hypothesis was that each of variable would peak at intermediate deer

browse intensity, in accordance with the intermediate disturbance hypothesis (IDH) (Connell,

1978). These results would provide more support for the IDH and justifies further research on

how deer have altered forest’s canopy structural complexity and function throughout the U.S.

Anticosti Island:

During the late 19th century a herd of nearly 200 deer were imported to Anticosti Island,

Quebec by chocolatier Henri Menier. Anticosti Island serves as a 7,493 km2 enclosure where

game are unable to migrate due to the island’s distance from the mainland. Therefore, the deer

have proliferated throughout the island and have rapidly altered the forest landscape. Balsam fir

stands were formerly the dominant forest type, but have shifted to white spruce with intense

herbivory (Potvin et al, 2003). Remaining balsam fir stands are expected to be eliminated within

30 to 40 years (Potvin et al. 2003).

In this study, we used a chronosequence of disturbance to investigate how deer have

influenced canopy structure, biomass, and species diversity over the last 120 years on Anticosti

Island. We investigated how these variables change along a chronosequence of 30, 50, 70, 90,

and 120 year old plots within pure white spruce or balsam fir stands to represent the former and

current dominant forest types (n=14) (Figure 2).

Our null hypothesis is that both time and deer browse will have no effect on the structural

complexity and biomass of both white spruce and balsam fir. Our alternative hypothesis is that

there will be an increase in structural complexity and biomass over time and a significant

difference between the structural complexities of both balsam fir and white spruce stands at the

70 and 90 year age class. We could only compare white spruce and balsam fir at 70 and 90 year

Reed 6

old plots because these are the only age classes where both exist, since deer had not spread

across the entire island until after 70 years. We predicted that both rugosity and biomass would

have a positive correlation with time due to both variables being dependent on stand height and a

well-developed canopy. We predicted a difference in structural complexity between balsam fir

and white spruce due to a predicted lack of a mid-story within balsam fir stands, as intense

browse pressure removed regenerating balsam fir seedlings. These results would indicate that

deer are having a significant effect on the canopy structural complexity of Anticosti’s forests and

that they could be influencing the carbon sequestering ability of the landscape.

Literature Review:

Forest Disturbance:

Although “ecological disturbance” often has a negative connotation, both abiotic and

biotic disturbances are crucial to the function and structure of forests worldwide. Within the

United States, forest dynamics are driven by herbivory, fire, drought, introduced species, insect

outbreaks, hurricanes, or ice storms, with each of these disturbance regimes being altered by

climate change and anthropogenic forcings (Dale et al. 2001). The intensity of these disturbances

can have extensive effects on the development and function of forest ecosystems over both brief

and extended time scales.

One of ecological disturbance’s influences over long time periods is succession, or the

changes in a community when a perturbation occurs (Connell, 1977). Disturbance facilitates

forest succession by opening gaps within the canopy, which releases tree species residing in the

understory. Over time, frequent perturbations create a developed forest with multiple age classes

and higher species diversity. Forests altered by disturbances can then have increased or

decreased function, or the “work” that an ecosystem does, such as cycling nutrients or

Reed 7

maintaining wildlife habitat. For example, fire drives forest dynamics throughout most of North

America through seedling regeneration, nutrient cycling, and improved wildlife habitat (Abrams

et al. 1992; Russell et al. 2009). These benefits to forests are then conferred in different ways

through other disturbances.

In addition, disturbance intensity is intimately tied to ecosystem stability, specifically

ecological resilience and resistance. Resilience is defined as the rate that an ecosystem returns to

a pre-disturbance functional state, but this rate is directly proportional to the intensity of

ecosystem disturbance (Attiwill, 1994). Resistance is an ecosystem’s “inertia,” or the system’s

ability to absorb disturbance without large deviations from the current functional state (Attiwill,

1994). These measures of stability are believed to be directly proportional to the level of

diversity within an ecosystem, known as the diversity-stability hypothesis (MacArthur, 1955;

Tilman and Downing, 1994). Nevertheless, this hypothesis appears to be largely dependent on

controlled experiments, where the community assemblages are random and only one trophic

level is analyzed (Hooper et al. 2005). In addition, ecosystem function within many diversity-

resilience studies is characterized by biomass production within grasslands (Tilman et al. 1994,

2001; Hector et al. 1999) while many other equally important functions and ecosystems remain

unaccounted for.

The diversity-resilience/resistance hypothesis lacks an incorporation of ecosystem

stochasticity, as a wide variety of top-down and bottom-up disturbances are not considered.

Forest ecosystems are dynamic in that they can and will respond in many ways to the wide

variety of disturbances that affect them. To account for this, the intermediate disturbance

hypothesis (IDH) focuses on varying disturbance frequency and intensity, predicting that

biodiversity and function will peak at intermediate levels of disturbance (Connell, 1978). The

Reed 8

mechanism behind IDH is that mid-frequent disturbances provide time for new species to

establish between perturbation bouts, while eliminating competitive exclusion by reducing the

amount of species that could dominant a landscape (Connell, 1978). This hypothesis depends on

a non-equilibrium state within a community and can be applied to a multitude of disturbances.

Most studies only consider IDH through the lens of species diversity, while there are

other ecological indicators that the hypothesis can be applied to. Fahey et al. (2015) found that

moderate wind disturbance improved canopy structural complexity and potentially carbon

sequestration. In addition, Biswas and Mallik (2010) found that functional traits such as

productivity and disturbance tolerance increased with moderate disturbance. Nevertheless, IDH

is highly contested on the grounds that fewer than 20% of studies have found the predicted peak

in diversity or function at moderate disturbance and that there are logical inconsistencies within

the hypothesis (Fox, 2013).

Materials and Methods:

Sampling:

Between both Anticosti and Allegheny, we established 3, 30 x 5 m belt transects in each

plot and recorded canopy complexity, species composition, and diameter at breast height (DBH)

for all trees > 5 cm DBH. Canopy complexity, or rugosity, was measured using the ground-based

portable canopy LiDAR and metrics of complexity were calculated following the methods of

Hardiman et al. (2011). Transect length was roughly double the height of the canopy and transect

width captured most branches hanging above the LiDAR.

Reed 9

Portable Canopy LiDAR:

Canopy structural complexity was measured using the ground-based portable canopy

LiDAR system (PCL; Parker et al. 2004a). The PCL measures the arrangement of leaves and

stems within a canopy using an upward-facing infrared laser at 100 Hz along a transect. Utilizing

a REIGL rangefinder and LaserWin software we recorded 7500-8500 laser returns across each

transect to collect a representative and consistent sample. These returns were then processed in

MatLab, which produced key structural complexity metrics, such as rugosity, porosity, and

canopy height (Table 1). In addition, we produced hit grids to illustrate the density of returns

across each transect in 1 m x 1 m bins that extend in both the horizontal and vertical directions

(Figure 1).

Allegheny Plateau

Study Site:

This experiment took place on the Allegheny Plateau in coordination with the U.S. Forest

Service. We used four sites in north-western (Fool’s Creek, Deadman’s Corners) and north-

central (Gameland 30, Wildwood Tower) Pennsylvania on the unglaciated section of the

Appalachian Plateau (Horsley et al. 2003) (Table 2). The landscape is composed of contiguous

forest with interspersed agricultural and natural gas extraction land uses. Precipitation ranges

from 1020 to 1070 mm per year and average temperature ranges from 8 to 9 °C (McNab and

Avers, 1994). The soils are highly acidic, with the dominant soil orders being alfisols, entisols,

inceptisols, and ultisols made from parent materials of sandstone, siltstone, and shale (McNab

and Avers, 1994; Horsley et al. 2003). Two sites, Fool’s Creek and Deadman’s Corners, were

located within the Allegheny National Forest, one within State Gamelands 30, and one

(Wildwood Tower) on land belonging to the National Gas Fuel Company. All sites were

Reed 10

previously composed of 60-70 year old stands of black cherry, red maple, and sugar maple

(Tilghman, 1989).

Deer Enclosure:

Within each site, a 65 ha deer enclosure with 2.5 m high fencing was assembled. Each

enclosure experimentally manipulated deer populations at different densities of 4, 8, 15, and 25

deer/km2 for 10 years (Tilghman, 1989; Horsley et al. 2003; Nuttle et al, 2011). Two enclosures

were established in 1979 and the other two in 1980. Each density was treated with three different

cutting schemes (clear cut, cut to 60% residual relative density, and uncut). There have never

been any measurements of canopy structural complexity within these plots.

Transect and Plot Selection:

We measured a total of 16 plots (n=16). Within each plot, the first transect was randomly

established by selecting a GPS point in the plot’s center prior to entering. From this central point,

two other parallel transects were established 30 m away from the original. We only collected

metrics of structural complexity within each subenclosure’s clear-cut treatment.

Statistical Tests:

To calculate biomass I used Jenkins et al. (2003) allometric equations to determine the

biomass per species within each plot from the DBH measurements. From these calculations we

used RStudio’s vegan package to calculate species richness, evenness, and Shannon diversity for

each plot. From these data I used linear regressions to determine how biomass, canopy structural

complexity, and Shannon diversity change with increasing deer browse. In addition, I used One

Way ANOVAs to find an overall effect of deer density and Tukey’s t-test to compare the effects

Reed 11

of each of the density treatments within the plots. SigmaPlot was the primary statistical software

used to calculate these metrics.

Anticosti Island:

Study Site:

This experiment occurred on Anticosti Island (7943 km2) in the Gulf of St. Lawrence,

Quebec (49°50’19.000”N, 64°17’36.000”W). The climate on the island is maritime and average

temperatures are -13.6 C in January and 14.8 C in July (Potvin, 2003). The forests on Anticosti

are within the boreal zone and belong to the eastern balsam fir-white birch climactic region

(Saucier et al. 2003).

From 1896 to 1897 nearly 220 deer were released on the island, with no large predators

to suppress the population (Tremblay et al. 2006). Their population peaked within 30 years and

has remained consistent at 20 deer/km2 over the last 130 years (Potvin and Breton, 2005). Intense

browse preference for balsam fir has led to its lower regenerative ability and increased white

spruce populations (Potvin et al. 2003). Since deer are unable to leave or enter Anticosti, the

island serves as a natural enclosure and provides insight to the effects of deer on boreal systems.

Plot Selection:

On Anticosti Island we measured 16 plots. Each age class contained two replicates and

each was at least 10 km away from their respective replicate. Within each plot, the first transect

was randomly established by selecting a GPS point in the plot’s center prior to entering. From

this central point, two other parallel transects were established 30 m away from the original.

Finally, all stands were similar density, as designated by the natural resource group on the island.

This density was the most prevalent on the island and is the most representative.

Reed 12

Statistical Test:

I calculated biomass using the fir and spruce allometric equations provided by Jenkins et

al. (2003). I then analyzed changes in white spruce rugosity and biomass over time with linear

and quadratic regressions to find an overall effect of age. SigmaPlot was the primary statistical

software used to calculate these metrics.

Results:

Allegheny Plateau:

My findings indicate a significant increase in canopy structural complexity in the 15

deer/km2 treatment, as average rugosity increased from roughly 6.58 m (8 deer/km2) to 10.15 m

(15 deer/km2) (p<0.05; Figure 2; Table 4; Table 5). Between all other comparisons of structural

complexity and deer density, there were no statistically significant differences between other

treatments (Appendix; Table A)

Biomass and diversity had more variable responses to deer density than canopy structural

complexity. There was a significant decrease in biomass with increasing deer density, especially

from 4 and 15 deer/km2 to 25 deer/km2 (r2=0.294; p<0.001; Figure 3; Table 6; Table 7). Biomass

dropped significantly from the 4 and 15 deer/km2 treatments to the 8 deer/km2 treatments

(p=0.001, p=0.002; Table 6; Table 7) while there was no substantial difference in biomass from

the 8 and 25 deer/km2 treatments (p>0.05). Shannon Diversity also decreased significantly with

increasing deer density, but there were no significant differences between each deer density

treatment (r2=0.199; p<0.0001; Figure 4; Table 8; Table 9).

Anticosti Island:

Reed 13

On Anticosti, white spruce stand rugosity increased significantly as time passed

(r2=0.791; p<0.0031; Figure 5; Table 10). Following 70 years, there is no noticeable change in

rugosity and the curve stabilizes. With regard to balsam fir stands, there are no significant

differences in rugosity between the 70, 90, and 120 year age classes and no differences between

balsam fir and white spruce rugosity at the 70 and 90 year age classes (p>0.05).

For all four white spruce age classes, biomass increases regularly until the 90 year age

class, after which it decreases (r2 = 0.777; p < 0.05; Figure 6; Table 11). The pattern that emerges

is hyperbolic and peaks at the 70 year age class. In addition, there is a strong relationship

between biomass and rugosity, wherein rugosity increases with white spruce biomass until the 70

year age class (r2=0.769; p=0.0216; Figure 7; Table 12). In contrast, there was no pattern in

balsam fir biomass from 70 to 120 years old.

Discussion:

Allegheny Plateau:

The increase in rugosity from 8 deer/km2 to 15 deer/km2 was originally believed to be an

intermediate disturbance effect (Figure 3), but there is no significant increase in species diversity

at moderate deer browse intensity (Figure 4), as would be predicted by the IDH. An alternative

explanation is that the significant increase in biomass from 8 deer/km2 to 15 deer/km2 (p=0.002;

Table 4) caused an increase in structural complexity. This increase in structural complexity and

biomass may have been due to natural pin cherry mortality, since this species has a 30-40 year

lifespan (Burns and Honkala, 1990) and was abundant in the 8 deer/km2 treatment. There was

more pin cherry located within the 8 deer/km2 treatment, due to competition from black cherry

stump sprouts within the 4 deer/km2 treatments and intense browse preference in the 15 deer/km2

Reed 14

treatment (Ristau and Horsley, 1999; Tilghman, 1989). As pin cherry died within the 8 deer/km2

treatment, there were likely increased canopy gaps and less living material within each plot,

decreasing both rugosity and biomass (Table 3; Table 7). These data indicate that IDH may not

explain the changes in structural complexity and that our results could be due to tree mortality

and new gap formation within the 8 deer/km2 treatment. In addition, these results may indicate

that rugosity is the most precise measure of structural complexity, as it was the only PCL metric

that recorded changes within the canopy.

Further, if the biomass at the 8 deer/km2 treatment decreased recently due to pin cherry

mortality, it is possible that the biomass at this treatment is similar to that of 4 and 15 deer/km2.

If this is the case, the sampled forests may have reached a threshold of deer browse disturbance

at the 15 deer/km2 treatment, past which ecosystem function rapidly declines (p<0.001; Figure

3). Similar non-linear responses have been observed in other forests, wherein there were abrupt

declines in ANPP when basal area loss was above either 61% or 66% (Stuart-Haentjens et al.

2015).

These results do not support my original hypothesis of the intermediate disturbance effect

influencing canopy structural complexity, biomass, or biodiversity; therefore, we accept the null

hypothesis. In contrast, we found that rugosity is likely the best indicator of structural complexity

in light of deer browse and that intense deer browse may lead to a decline in biomass once a

certain intensity threshold is reached. These results are relevant due to the high abundance of

deer throughout the eastern United States and the lack of understanding of how ungulates alter

ecosystem function and health. We believe these data could inform both land managers and

citizens as to the proper deer stocking level within an ecosystem in order to maximize function,

health, and recreation.

Reed 15

Anticosti Island:

The research on Anticosti Island was entirely experimental, in that it was the first time

the PCL had been used within a boreal ecosystem. I found an increase in white spruce structural

complexity with an increase in age from 30 to 70 years (p<0.1; Figure 5). In contrast, there were

no discernable changes in structural complexity for balsam fir over time or differences in

structural complexity between balsam fir and white spruce in the 70 and 90 year age classes.

Therefore, I can partially reject the null hypothesis and conclude that white spruce forests

increase in rugosity over time.

The pattern of white spruce stand rugosity over a chronosequence is similar to the

relationship Hardiman et al. (2013b) found in northern lower Michigan, where canopy rugosity

increases with forest age. This similarity indicates that the PCL can be used within boreal forests,

even though rugosity values are low in comparison to deciduous forests. In addition, rugosity

values may serve as an accurate predictor of biomass and carbon sequestration across boreal

landscapes. These results illustrate the need to better understand boreal canopies and how their

arrangement can influence the amount of carbon absorbed by a system.

Conclusions:

Over the last 100 years, Allegheny and Anticosti’s intense deer herbivory has shifted the

dominant forest species and caused significant deviations in canopy structural complexity and

biomass. Within Allegheny, our results indicate that there may be a threshold of disturbance for

northern hardwood forests – even after 30 years. This helps to inform what the ideal stocking

capacity could be within the region and contributes to our understanding of tipping points within

ecosystems. In Anticosti, the rise in canopy structural complexity and biomass indicates that the

PCL may be a useful tool in boreal systems and could serve as a predictor of carbon

Reed 16

sequestration or wildlife habitat. Although no direct comparisons could be made between

Allegheny and Anticosti, deer browse has significantly impacted these large landscapes and

those living on them. These understudied, widespread, and long-term consequences of herbivory

indicate that we have an unsatisfactory understanding of how deer influence eastern forest’s

functional and adaptive capacity in light of increasing disturbance through climactic and

anthropogenic forcings.

Acknowledgements:

I thank Peter Curtis, Roger Williams, Lauren Pintor, Alex Fotis, Maggie Woodworth, Nicolas

Houde, Kathleen Knight, Charlie Flower, Alex Royo, Todd Ristau, JP Tremblay

In addition, I thank the U.S. Forest Service Northern Research Station, The University of Laval,

OARDC: SEED Fellowship, and the Undergraduate Research Office Summer Fellowship

Reed 17

Works Cited:

Abrams, M. D. 1992. Fire and the development of oak forests – In eastern North America, oak

distribution reflects a variety of ecological paths and disturbance conditions.

Bioscience 42:346-353.

Attiwill, P. M. 1994. The disturbance of forest ecosystems – The ecological basis for

conservative management. Forest Ecology and Management 63:247-300.

Bengtesson, J., S. G. Nilsson, A. Franc, and P. Menozzi. 2000. Biodiversity, disturbances,

ecosystem function and management of European forests. Forest Ecology and

Management 132:39-50.

Biswas, S. R., and A. U. Mallik. 2010. Disturbance effects on species diversity and functional

diversity in riparian and upland plant communities. Ecology 91:28-35.

Burns, R. M., and B. H. Honkala. 1990. Silvics of North America: 1. Conifers; 2. Hardwoods.

Agriculture Handbook 654. U.S. Department of Agriculture, Forest Service, Washington,

DC. 2; 877

Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs – High diversity of trees and

corals is maintained only in a non-equilibrium state. Science 199:1302-1310.

Connell, J. H., and R. O. Slatyer. 1977. Mechanisms of succession in natural communities and

their role in community stability and organization. American Naturalist 111:1119-1144.

Cote, S. D., T. P. Rooney, J. P. Tremblay, C. Dussault, and D. M. Waller. 2004. Ecological

impacts of deer overabundance. Annual Review of Ecology Evolution and

Systematics 35:113-147.

Dale, V. H., L. A. Joyce, S. McNulty, R. P. Neilson, M. P. Ayeares, M. D. Flannigan, P. J.

Hanson, L. C. Irland, A. E. Lugo, C. J. Peterson, D. Simberloff, F. J. Swanson, B. J.

Stocks, and B. M. Wotton. 2001. Climate change and forest disturbances.

Bioscience 51:723-734.

Fox, J. W. 2013. The intermediate disturbance hypothesis should be abandoned. Trends in

Ecology & Evolution 28:86-92.

Goetsch, C., J. Wigg, A. A. Royo, T. Ristau, and W. P. Carson. 2011. Chronic over browsing

and biodiversity collapse in a forest understory in Pennsylvania: Results from a 60 year-

old deer exclusion plot. Journal of the Torrey Botanical Society 138:220-224.

Hardiman, B. S., G. Bohrer, C. M. Gough, and P. S. Curtis. 2013a. Canopy Structural Changes

Following Widespread Mortality of Canopy Dominant Trees. Forests 4:537-552.

Reed 18

Hardiman, B. S., C. M. Gough, A. Halperin, K. L. Hofmeister, L. E. Nave, G. Bohrer, and P. S.

Curtis. 2013b. Maintaining high rates of carbon storage in old forests: A mechanism

linking canopy structure to forest function. Forest Ecology and Management 298:111-

119.

Hector, A., B. Schmid, C. Beierkuhnlein, M. C. Caldeira, M. Diemer, P. G. Dimitrakopoulos, J.

A. Finn, H. Freitas, P. S. Giller, J. Good, R. Harris, P. Hogberg, K. Huss-Danell, J. Joshi,

A. Jumpponen, C. Korner, P. W. Leadley, M. Loreau, A. Minns, C. P. H. Mulder, G.

O'Donovan, S. J. Otway, J. S. Pereira, A. Prinz, D. J. Read, M. Scherer-Lorenzen, E. D.

Schulze, A. S. D. Siamantziouras, E. M. Spehn, A. C. Terry, A. Y. Troumbis, F. I.

Woodward, S. Yachi, and J. H. Lawton. 1999. Plant diversity and productivity

experiments in European grasslands. Science 286:1123-1127.

Hooper, D. U., F. S. Chapin, J. J. Ewel, A. Hector, P. Inchausti, S. Lavorel, J. H. Lawton, D. M.

Lodge, M. Loreau, S. Naeem, B. Schmid, H. Setala, A. J. Symstad, J. Vandermeer, and

D. A. Wardle. 2005. Effects of biodiversity on ecosystem functioning: A consensus of

current knowledge. Ecological Monographs 75:3-35.

Horsley, S. B., S. L. Stout, and D. S. DeCalesta. 2003. White-tailed deer impact on the

vegetation dynamics of a northern hardwood forest. Ecological Applications 13:98-118.

Ishii, H. T., S. Tanabe, and T. Hiura. 2004. Exploring the relationships among canopy structure,

stand productivity, and biodiversity of temperature forest ecosystems. Forest

Science 50:342-355.

Jenkins, J. C., D. C. Chojnacky, L. S. Heath, and R. A. Birdsey. 2003. National-scale biomass

estimators for United States tree species. Forest Science 49:12-35.

Jung, K., S. Kaiser, S. Bohm, J. Nieschulze, and E. K. V. Kalko. 2012. Moving in three

dimensions: effects of structural complexity on occurrence and activity of insectivorous

bats in managed forest stands. Journal of Applied Ecology 49:523-531.

Leopold, A., L. K. Sowls, and D. L. Spencer. 1947. A survey of over-populated deer ranges in

the United States. Journal of Wildlife Management 11:162-177.

Macarthur, R. 1955. Fluctuations of animal populations, and a measure of community stability.

Ecology 36:533-536.

McNab, W. H., and P. E. Avers, compilers. 1994. Ecological subregions of the United States:

section descriptions. USDA Forest Service Administrative Publication WO- WSA-5.

USDA Forest Service, Washington, D.C., USA.

Nuttle, T., E. H. Yerger, S. H. Stoleson, and T. E. Ristau. 2011. Legacy of top-down herbivore

pressure ricochets back up multiple trophic levels in forest canopies over 30 years.

Ecosphere 2.

Reed 19

Parker, G. G., D. J. Harding, and M. L. Berger. 2004. A portable LIDAR system for rapid

determination of forest canopy structure. Journal of Applied Ecology 41:755-767.

Potvin, F., P. Beaupre, and G. Laprise. 2003. The eradication of balsam fir stands by white-tailed

deer on Anticosti Island, Quebec: A 150-year process. Ecoscience 10:487-495.

Potvin, F., and L. Breton. 2005. From the field: Testing 2 aerial survey techniques on deer in

fenced enclosures - visual double-counts and thermal infrared sensing. Wildlife Society

Bulletin 33:317-325.

Ristau, T. E., and S. B. Horsley. 1999. Pin cherry effects on Allegheny hardwood stand

development. Canadian Journal of Forest Research-Revue Canadienne De Recherche

Forestiere 29:73-84.

Russell, R. E., J. A. Royle, V. A. Saab, J. F. Lehmkuhl, W. M. Block, and J. R. Sauer. 2009.

Modeling the effects of environmental disturbance on wildlife communities: avian

responses to prescribed fire. Ecological Applications 19:1253-1263.

Saucier, J.-P., Grondin, P., Robitaille, A. & Bergeron, J.-F. (2003) Vegetation Zones and

Bioclimatic Domains in Québec. Ministère des Ressources naturelles et de la Faune,

Québec, Canada.

Stuart-Haentjens, E. J., P. S. Curtis, R. T. Fahey, C. S. Vogel, and C. M. Gough. 2015. Net

primary production of a temperate deciduous forest exhibits a threshold response to

increasing disturbance severity. Ecology 96:2478-2487.

Tilghman, N. G. 1989. Impacts of white-tailed deer on forest regeneration in northwestern.

Journal of Wildlife Management 53:524-532.

Tilman, D., and J. A. Downing. 1994. Biodiversity and stability in grasslands. Nature 367:363-

365.

Tilman, D., P. B. Reich, J. Knops, D. Wedin, T. Mielke, and C. Lehman. 2001. Diversity and

productivity in a long-term grassland experiment. Science 294:843-845.

Tilman, D., P. B. Reich, and J. M. H. Knops. 2006. Biodiversity and ecosystem stability in a

decade-long grassland experiment. Nature 441:629-632.

Tremblay, J.-P., J. Huot, and F. Potvin. 2006. Divergent nonlinear responses of the boreal forest

field layer along an experimental gradient of deer densities. Oecologia 150:78-88.

Reed 20

Tables:

Term Definition

Rugosity

The three-dimensional arrangement of leaves and stems within

the canopy

Height Variability of mean return height

Mode Variability of mean return height

ModeEl Mean height of maximum return density

MeanHeight Mean return height

MeanStd Mean variability of return height

MeanLAI Leaf Area Index (LAI) calculated from return

MeanTopel_CPAll Mean maximum height

Top_rug Variability of outer canopy surface height

Porosity Ratio of bins with no returns to total number of bins

Table 1: A list of metrics generated by the Portable Canopy LiDAR. Within both

the Allegheny Plateau and Anticosti Island, rugosity served as the most descriptive

metric, while other descriptors of canopy structure served to be less useful.

Site Abbreviation

Deadman's Corners DMC

Wildwood Tower WWT

Gamelands 30 GL

Fool's Creek FC

Table 2: A list of the 4 replicate deer enclosure sites and their abbreviations. Each

65 ha enclosure was subdivided into 4 deer densities of 4, 8, 15, and 25 deer/km2

and were in use from 1979/1980 until 1990.

Reed 21

Allegheny Plateau Anticosti Island

Plots n = 16 n = 14

Treatments 4, 8, 15, 25 Deer/Km2 30, 50, 70, 90, 120 Year Age

Class

Selection Haphazard Haphazard

Transects/Plot 3 3

Transect Size 3 x 30 m 3 x 30 m

Trees Recorded >5 cm >5 cm

Deer Preferred Pin Cherry Balsam Fir

Deer Avoided Black Cherry White Spruce

Table 3: Between Allegheny and Anticosti, deer density and age class were the

independent variables while the resulting dominant trees, canopy structural

complexity, and biomass are the dependent variables. In Allegheny deer have

shifted the species composition from pin cherry to black cherry and in Anticosti

from balsam fir to white spruce. The number of transects per plot, transect size,

and minimum DBH were consistent between both study sites.

Reed 22

Rugosity Diff of Means t P

15 vs. 8 Deer/Km2 0.18 3.55 0.024

15 vs. 25 Deer/Km2 0.12 2.35 0.17

15 vs. 4 Deer/Km2 0.095 1.86 0.31

4 vs. 8 Deer/Km2 0.087 1.68 0.31

25 vs. 8 Deer/Km2 0.061 1.20 0.44

4 vs. 25 Deer/Km2 0.025 0.49 0.63

Table 4: Comparison of rugosity within each deer density treatment (One Way

ANOVA; Tukey’s t-test), wherein there were only significant deviations in canopy

rugosity between 15 and 8 deer/km2.

Rugosity (m) DMC WWT GL FC

4 Deer/Km2 6.63 8.54 8.83 8.27

8 Deer/Km2 6.51 7.23 5.89 6.73

15 Deer/Km2 12.1 8.26 12.0 8.35

25 Deer/Km2 6.50 6.41 7.94 9.92

Table 5: PCL-recorded canopy rugosity for each plot in order of ascending deer

density treatment, with the minimum value in 8 deer/km2 and the maximum value

in 15 deer/km2.

Reed 23

Biomass (kg) Diff of Means t p

4 vs. 25 Deer/Km2 2520 6.31 <0.001

15 vs. 25 Deer/Km2 2270 5.69 <0.001

4 vs. 8 Deer/Km2 2010 5.04 0.001

15 vs. 8 Deer/Km2 1770 4.42 0.002

8 vs. 25 Deer/Km2 505 1.27 0.41

4 vs. 15 Deer/Km2 247 0.62 0.55

Table 6: Comparison of biomass within each deer density treatment (One Way

ANOVA; Tukey’s t-test). There were no differences in biomass within two density

groups (p>0.05; 8 and 25 deer/km2; 4 and 15 deer/km2), but between these two

density groups there was significantly higher biomass in 4 and 15 deer/km2 in

comparison to 8 and 25 deer/km2 (p<0.005)

R Rsqr Adj Rsqr Standard Error of Estimate

0.543 0.294 0.244 1070

Std. Error t p

y0 513 14.6 <0.0001

a 33.6 -2.42 0.0299

Table 7: Linear regression indicating the relationship between biomass and deer

density, wherein there is a significant downward trend in biomass as deer densities

increase (r2 = 0.244; p<0.0001)

Reed 24

Biomass (kg) DMC WWT GL FC

4 Deer/km2 8349 7451 8106 6495

8 Deer/km2 5408 6290 5445 5212

15 Deer/km2 7649 7393 7617 6755

25 Deer/km2 5557 4512 5091 5174

Table 8: Recorded biomass for each plot in order of ascending deer density

treatment, with the minimum value in 25 deer/km2 and the maximum in 4

deer/km2.

R Rsqr Adj Rsqr Standard Error of Estimate

0.446 0.199 0.142 0.346

Std. Error t p

y0 0.165 8.73 <0.0001

a 0.0109 -1.86 0.0831

Table 9: Linear regression illustrating the relationship between Shannon Diversity

and deer density, where there was a weak negative correlation between Shannon

Diversity and deer density (r2 = 0.199; p < 0.0001). High variance at 15 and 25

deer/km2 likely reduced the significant downward trend.

Reed 25

R Rsqr Adj Rsqr Standard Error of Estimate

0.889 0.791 0.756 0.663

Std. Error t p y0 0.671 -0.647 0.542

a 0.0105 4.76 0.0031

Table 10: Quadratic regression illustrating the relationship between rugosity and

white spruce age class, wherein there was a strong positive increase in rugosity as

white spruce age class increase (r2 = 0.791; p < 0.0031).

R Rsqr Adj Rsqr Standard Error of Estimate

0.882 0.777 0.688 2420

Std. Error t p

y0 7070 -2.72 0.04

a 259 4.14 0.01

b 2.14 -4.01 0.01

Table 11: Quadratic regression indicating that biomass significantly peaks at the

70 year age class and then decreases towards the 90 year age class (r2 = 0.777; p <

0.05).

Reed 26

R Rsqr Adj Rsqr Standard Error of Estimate 0.877 0.770 0.712 0.744

Std. Error t p

y0 0.856 -0.817 0.460

a 7.85E-05 3.66 0.0216

Table 12: Linear regression output for the relationship between rugosity and white

spruce biomass at 30, 50, and 90 year age classes, where there is a strong positive

relationship between rugosity and white spruce biomass (r2 = 0.77; p < 0.0216).

Reed 27

Figures:

Figure 1: Hit grid for a high rugosity transect, wherein darker shading indicates a

higher amount of PCL returns. Rugosity is then calculated using σ(σ[VAI]z)x.

Reed 28

Figure 2: There was no overall trend in rugosity with increasing deer density, but

there was a significant increase in rugosity from 8 to 15 deer/km2 (p <0.05).

Reed 29

Figure 3: There was a strong negative correlation between deer density and

biomass across the plots (r2 = 0.244; p<0.0001). The 25 deer/km2 treatment caused

the most dramatic decrease in biomass.

Reed 30

Figure 4: There was a moderately strong negative correlation between deer density

and Shannon Diversity (r2 = 0.199; p < 0.0001). High variation within the 15 and

25 deer/km2 treatments could weaken this trend, but may indicate that intense

herbivory leads to multiple levels of species diversity and evenness.

Reed 31

Figure 5: There was a strong positive correlation between white spruce age class

and stand rugosity using a quadratic equation (r2 = 0.791; p < 0.0031), where the

rugosity plateaus at the 70 year age class

Reed 32

Figure 6: White spruce biomass peaked at the 70 year age class and rapidly fell in

the 90 year old stands, creating a parabolic trend (r2 = 0.777; p < 0.05)

Reed 33

Figure 7: White spruce biomass and rugosity were strongly and positively

correlated within the 30, 70, and 90 year age classes. The 90 year age class was

excluded due to a decrease in biomass and a plateau in rugosity.

Reed 34

Appendix:

Rugosity Height Mode ModeEl Height MeanStd LAI Topel_CPAll Top_rug Porosity

DMC1 6.64 4.28 4.36 8.31 6.91 8.44 7.64 11.44 4.33 0.57

WWT1 8.54 4.59 4.10 11.84 9.71 11.73 7.45 14.06 3.28 0.49

GL1 8.83 3.91 3.46 12.16 9.83 11.11 7.53 14.54 2.53 0.45

FC1 8.27 4.95 5.08 10.17 8.11 9.41 7.51 12.48 4.80 0.57

DMC2 6.51 4.24 4.39 9.12 7.52 6.88 6.92 11.26 4.28 0.55

WWT2 7.23 3.50 3.55 7.94 6.15 7.77 7.37 10.31 3.74 0.53

GL2 5.89 4.12 3.96 9.13 7.46 8.24 7.51 12.03 3.65 0.48

FC2 6.73 4.54 4.31 11.03 9.07 9.96 7.52 13.32 3.70 0.50

DMC3 12.05 5.21 5.43 10.04 7.88 12.40 7.84 12.66 5.10 0.56

WWT3 8.26 3.70 3.57 7.03 5.67 6.91 7.41 9.47 4.60 0.57

GL3 11.98 5.26 4.93 13.32 11.18 16.50 7.67 16.51 3.08 0.49

FC3 8.35 4.77 4.72 11.20 9.50 9.17 7.59 13.23 3.96 0.53

DMC4 6.50 4.06 3.95 9.32 7.64 8.18 7.30 11.51 3.27 0.52

WWT4 6.41 3.66 2.65 17.56 15.78 7.32 5.16 18.73 2.09 0.51

GL4 7.94 4.33 3.82 11.90 9.88 9.72 6.18 14.12 3.19 0.52

FC4 9.92 5.71 5.03 13.57 11.50 11.82 5.30 15.29 4.02 0.50

Table A: Aggregation of canopy structural complexity values across deer density treatments